Biosensors and Bioelectronics 155 (2020) 112111

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: http://www.elsevier.com/locate/bios

Improving the reproducibility, accuracy, and stability of an electrochemical

biosensor platform for point-of-care use
Lung-Chieh Chen a, b, 1, Erick Wang a, 1, Chun-San Tai a, b, 1, Yuan-Chen Chiu a, b, Chang-Wei Li a, c,
Yan-Ren Lin d, e, f, Tsung-Han Lee a, d, Ching-Wen Huang a, h, i, Jung-Chih Chen a, g, **, Wen
Liang Chen a, *
a
Department of Biological Science and Technology, National Chiao Tung University, Hsinchu, Taiwan
b
Institute of Molecular Medicine and Bioengineering, National Chiao Tung University, Hsinchu, Taiwan
c
AllBio Life Inc, Taichung, Taiwan
d
Department of Emergency Medicine, Changhua Christian Hospital, Changhua, Taiwan
e
School of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan
f
School of Medicine, Chung Shan Medical University, Taichung, Taiwan
g
Institute of Biomedical Engineering, National Chiao Tung University, Hsinchu, Taiwan
h
Department of Emergency, Shuang Ho Hospital, Taipei Medical University, New Taipei City, Taiwan
i
Division of Thoracic Surgery, Department of Surgery, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan

A R T I C L E I N F O A B S T R A C T

Keywords: Electrochemical biosensors possess numerous desirable qualities for target detection, such as portability and ease
Semiconductor manufacturing technology of use, and are often considered for point-of-care (POC) development. Label-free affinity electrochemical bio­
Biotin-streptavidin system sensors constructed with semiconductor manufacturing technology (SMT)-produced electrodes and a streptavi­
Label-free electrochemical platform
din biomediator currently display the highest reproducibility, accuracy, and stability in modern biosensors.
Point-of-care testing (POCT)
However, such biosensors still do not meet POC guidelines regarding these three characteristics. The purpose of
this research was to resolve the limitations in reproducibility and accuracy caused by problems with production
of the biosensors, with the aim of developing a platform capable of producing devices that exceed POC standards.
SMT production settings were optimized and bioreceptor immobilization was improved through the use of a
unique linker, producing a biosensor with exceptional reproducibility, impressive accuracy, and high stability.
Importantly, the three characteristics of the sensors produced using the proposed platform all meet POC stan­
dards set by the Clinical and Laboratory Standards Institute (CLSI). This suggests possible approval of the bio­
sensors for POC development. Furthermore, the detection range of the platform was demonstrated by
constructing biosensors capable of detecting common POC targets, including circulating tumor cells (CTCs),
DNA/RNA, and curcumin, and the devices were optimized for POC use. Overall, the platform developed in this
study shows high potential for production of POC biosensors.

1. Introduction manufacturing technology (SMT)-produced electrodes and a bio­

Electrochemical biosensors possess a low sample volume require­
ment, a short detection time, high portability, and ease of use (Anik
2017; da Silva et al., 2017; Szunerits et al., 2019), and are often the
foundation upon which various point-of-care (POC) products are
developed. Currently, the most promising design for POC use is a
label-free affinity electrochemical biosensor comprising semiconductor
synthesized streptavidin mediator. SMT enhances the consistency of
electrode assembly to improve reproducibility (Hierlemann et al., 2003;
Pohanka 2017). Label-free affinity detection offers ease of production
and direct interaction with a wide range of targets, which expands the
potential uses and improves the accuracy of the biosensors (Daniels and
Pourmand 2007; Johnson and Krauss 2017). Streptavidin exhibits strong
binding affinity for biotinylated bioreceptors and stabilizes bioreceptor

* Corresponding author. Department of Biological Science and Technology, National Chiao Tung University, Hsinchu, 30010, Taiwan, Republic of China.
** Corresponding author. Institute of Biomedical Engineering, National Chiao Tung University, Hsinchu, 30010, Taiwan, Republic of China.
E-mail address: wenurea@mail.nctu.edu.tw (W.L. Chen).
1
These authors contribute equally to this work.

https://doi.org/10.1016/j.bios.2020.112111
Received 30 September 2019; Received in revised form 12 February 2020; Accepted 18 February 2020
Available online 20 February 2020
0956-5663/© 2020 Elsevier B.V. All rights reserved.
L.-C. Chen et al.
modification to increase biosensor stability (Dundas et al., 2013). Thus,
biosensors produced in this way possess greater reproducibility, accu­
racy, and stability (three characteristics integral to POC products) than
those produced with other methods, as shown by a human immuno­
globulin A sensor designed by Minamiki et al., (2015). However, despite
these innovations, POC guidelines demand specifically high levels of
reproducibility, accuracy, and stability before biosensors can be
considered for POC use.
According to the Clinical and Laboratory Standards Institute (CLSI;
EP05-A3, EP24-A2, EP25-A), coefficient of variation (CV) less than 10%
for reproducibility, accuracy, and stability are required. Therefore,
further improvements to biosensors are necessary, specifically regarding
the electrodes and mediator, the two components that determine the
reproducibility, accuracy, and stability of all electrochemical biosensors.
In terms of electrode assembly, SMT production settings can be adjusted
to fit label-free affinity detection. This detection method measures
changes in electrical signal directly on the surface of the chip, and ac­
curacy is therefore highly dependent on electrode production settings.
The thickness of the thin-film metal and surface topography (roughness)
are two vital contributors to accuracy and can drastically affect con­
ductivity and consistency, respectively (Butterworth et al., 2019), and
ultimately affect the detection results. Consequently, calibration of the
SMT production settings presents the opportunity to considerably
improve accuracy. Meanwhile, despite the improvements in stability
achieved using a streptavidin biomediator, immobilization of the bio­
receptor directly onto the mediator limits its orientation and interferes
with function, thus limiting accuracy (Trilling et al., 2013). Introducing
a linker to the biomediator has been shown to be a feasible solution;
however, the optimal linker sequence to maximize accuracy remains
unclear and requires further research (Ikonomova et al., 2018).
The goal of this research was to develop a semiconductor
manufacturing electrochemical biosensor (SMEB) platform that com­
bines improved SMT production settings with a unique streptavidin
biomediator to produce label-free affinity electrochemical biosensors
suitable for POC use. The thickness and roughness of the electrodes were
calibrated to optimize the reproducibility of label-free affinity detection,
while a unique linker with ideal flexibility as well as rigidity was added
to the streptavidin biomediator to improve accuracy. To prove func­
tionality, a finished sensor was tested using protein samples to confirm
reproducibility, accuracy, and stability. After verification of perfor­
mance, different biosensors were constructed to detect other common
POC targets, including circulating tumor cells (CTCs), DNA/RNA, and
small compounds (e.g. curcumin), and their functionality was tested.
Additionally, to optimize the platform for future development of a POC
diagnosis kit, the guidelines for POC product development (Siontorou
and Batzias 2010) were followed. This involved collaborating with
manufacturers and making improvements, including modification of
standard operating procedure (SOP) definitions, optimization of reac­
tion conditions, simplification of measurement technology, and cus­
tomization of the detection machine. Overall, the SMEB platform shows
strong potential for integration in POC products for various purposes,
such as food safety monitoring, environmental monitoring, disease di­
agnostics, and clinical guidance.
2. Materials and methods
2.1. Materials and apparatus
The pET-30a(þ) vector was purchased from Novagen, Inc., and iso­
propyl β-D-1-thiogalactopyranoside (IPTG), mercaptoundecanoic acid
(11-MUA), N-ethyl-N’-(3-dimethylaminopropyl) carbodiimide hydro­
chloride (EDC), N-hydroxysuccinimide (NHS), and curcumin powder
were purchased from Sigma Aldrich, Inc. Ni-NTA resin was ordered from
BioSmart, Inc. and an NHS-PEG4-biotinylation kit was purchased from
Thermo Fisher Scientific, Inc. Anti-cardiac troponin I (cTnI) antibody
was ordered from Proteintech, Inc. cTnI native protein was ordered from
2
Biosensors and Bioelectronics 155 (2020) 112111
Mybiosource, Inc. Anti-cytokeratin 7 (CK7) antibody was purchased
from GeneTex, Inc. All aqueous solutions were prepared using ultrapure
water from a Milli-Q Direct machine, and all oligonucleotides (DNA
probe and DNA fragments) were synthesized by Tri-I Biotech, Inc.
The surface topography of the bio-chips was examined using a
scanning electron microscope (SEM; Hitachi SU8010) at an accelerating
voltage of 15 keV, and all electrochemical measurements were per­
formed using a CH Instrument 920D (CHI 920D; CH instruments, Austin,
TX, USA). Development of the final customized device was achieved
through cooperation with Zensor R&D, Inc.
2.2. Electrode production
Gold thin-film electrodes were produced using direct plate copper
(DPC) technology via a sputtering and etching process, illustrated in
Fig. 1A, which is identical to SMT electrode production (Hierlemann
et al., 2003). In this study, electrode settings were adjusted to a thickness
greater than 0.1 μm and a surface roughness less than 0.3 μm. To eval­
uate the quality of production, the electrodes were examined using X-ray
fluorescence spectroscopy and an alpha-step profiler; the results are
shown in Fig. 1B.
2.3. Biomediator production
For the biomediator, two types of linker, GS linker (G4S1) and GW
linker (GINSSSVPGDPPW) (Lower et al., 2005), were independently
constructed with the streptavidin sequence and were cloned into the
pET-30a(þ) vector for recombinant protein expression. Escherichia coli
(E. coli) DH5α was selected as the host for plasmid cloning and propa­
gation (Fig. S1A), while E. coli BL21 (DE3) was used as the host for re­
combinant protein expression. Protein expression was regulated by IPTG
induction and protein purification was executed through a standard
Ni-NTA column protocol (Fig. S1B).
2.4. SMEB platform fabrication
Establishment of the SMEB platform consisted of five steps from self-
assembled monolayer (SAM) formation to bioreceptor modification
(Fig. 2A). First, SAMs were established on the gold surface by immersing
the electrodes in an ethanolic solution of 11-MUA (Stettner et al., 2009).
The carboxyl-terminated group of 11-MUA was then activated through
the EDC and NHS coupling reaction (Palazon et al., 2014; Tsai et al.,
2019). The electrodes were rinsed and immersed in the biomediator
protein solution (conc. 10 μg/mL). After the biomediator was immobi­
lized on the gold surface by covalent bonding, the electrodes were
washed and blocked with 3% gelatin solution. To immobilize the bio­
receptors (antibody, DNA probe and functional protein) on the elec­
trodes, NHS-PEG4-biotinylation was performed and biotinylated
bioreceptors were added to the surface of the biomediator-modified
electrodes. Finally, the electrodes were washed and dried, and electro­
chemical impedance spectroscopy (EIS) was performed to verify the
entire process of SMEB platform fabrication (Fig. 2B). Electrodes were
stored at 4 � C for subsequent use.
2.5. Electrochemical measurements
For EIS measurements, a frequency range of 0.1–1 kHz with an
amplitude of 10 mV was set to execute scanning, ZSimpWin was used to
analyze the EIS data, and Randles’ equivalent circuit R(Q(RW)) was
applied to simulate the value of the charge transfer resistance (Rct) as a
sample detection value. For differential pulse voltammetry (DPV)
measurements, a potential range of 0.6 V to -0.6 V and amplitude of 50
mV were selected to perform scanning and the sample detection value
was acquired from the absolute value of the maximum detected current.
All EIS and DPV values were normalized by subtracting the blank values
obtained from measuring the solution without target analytes. Based on
L.-C. Chen et al. Biosensors and Bioelectronics 155 (2020) 112111

Fig. 1. Production of electrodes. (A) Schematic illustration of electrode fabrication. (B) Production setting confirmation using X-ray fluorescence spectroscopy for
thickness (μm) determination and an Alpha-Step® D-500 Stylus Profiler for surface roughness (μm) evaluation. Spec.: specification; USL: upper specification limit;
LSL: lower specification limit; No.: chip number; Max.: maximum; Min.: minimum; Avg.: average; Std.: standard deviation.

Fig. 2. SMEB platform fabrication. (A) Schematic illustration of SMEB platform fabrication. (B) EIS measurements to verify the steps of antibody (a), DNA probe (b)
and αS1-casein (c) platform fabrication.

the obtained results (Fig. S2), EIS was selected for protein and CTC
detection, and DPV was selected for detection of DNA/RNA and small
compounds.
2.6. Calculation of reproducibility, accuracy, and stability
Accuracy describes the variation between biosensor-detected values
and the actual values. In tests of unknown samples shown in Fig. 4,
realistic values were determined by the gold standard method in the
hospital. The accuracy values of SMEB detection results were calculated
by equation (2) and are shown in Table 1.

Accuracy ¼ 100% jRelative ​ deviation ðRDÞj ¼ 100%

Reproducibility was calculated based on the coefficient of variation � �
(CV; equation (1)) value of the Rct values measured from each chip after �Detection value Realistic value �
�
� � 100% �
� (2)
production to represent the consistency of production and signal Realistic value
detection. Stability describes the variation in detection signals during a period
Standard deviation ðSDÞ of long-term storage. As shown in Fig. 5, the baseline of each analyte
Coefficient ​ of ​ variation ðCVÞ ¼
average of detection value ðSÞ
� 100% concentration was determined by the average of the values detected
every week and RD of the detection values compared to the baseline was
(1)
calculated to deduce the stability value by equation (3).

3
L.-C. Chen et al. Biosensors and Bioelectronics 155 (2020) 112111

Stability ¼ 100% jRelative ​ deviation ðRDÞj ¼ 100% Table 1

� � cTnI detection in whole blood samples.
�Week detection value ðWÞ baseline ðWÞ �
�
� � 100%�� (3) Sample no. SMEB (ng/mL) Hospital (ng/mL) Accuracy (%)
baseline ðWÞ
1 0.00 N.D. -
2 0.00 N.D. -
3. Results
3 0.02 N.D. -
4 0.00 N.D. -
To improve biosensor performance and fulfill the POC guidelines set 5 1.74 1.64 93.91
by the CLSI, SMT production settings were calibrated to produce elec­ 6 4.28 4.02 93.54
trodes with a thickness greater than 0.1 μm and a surface roughness less 7 0.35 0.37 94.60
8 0.72 0.71 98.60
than 0.3 μm to complement label-free affinity detection (Fig. 1). A GW 9 0.20 0.19 94.74
linker was also fused to the streptavidin biomediator to optimize flexi­ 10 0.62 0.64 96.88
bility and rigidity (Fig. 2), and thereby improve function. Functionality 11 3.19 3.08 96.43
was evaluated using cardiac troponin I (cTnI) to confirm that repro­
ducibility, accuracy, and stability all had CV of less than 10%.
After confirming the consistency of reproduction across all stages of
the SMEB platform (Fig. 3A-C), storage stability was examined by stor­
3.1. Production consistency and detection reproducibility
ing a batch of the electrodes at room temperature (RT) in a drier and
investigating their function over a period of nine weeks. Consistent
Robust electrode consistency is essential to improve biosensor
maintenance of the signal was achieved over time (Fig. 5); the signal
reproducibility. EIS was therefore performed on five chips produced variation over nine weeks was limited to 6%. This suggests stability of
from the same batch after electrode assembly and the CV was calculated the SMEB platform is 94%, which is appropriate for POC use.
to assess consistency. The calibrated settings improved the CV from
15.92% for the chips produced with original settings to 3.03% (Fig. 3A 3.3. Applications
and SEM images), and thus improved consistency. Next, the effects of
the calibrated settings, as well as the improved biomediator (with GW To demonstrate the capabilities of the completed SMEB platform,
linker), on modified chip consistency were investigated by performing three different biosensors were constructed (Fig. 2B) to detect circu­
EIS on four different sets of five modified chips, as shown in Fig. 3B and lating tumor cells (CTCs), DNA/RNA, and curcumin (Fig. S3).
the respective SEM images. In chips produced with the original SMT
production settings, changing the original biomediator (Fig. 3Ba) to the 3.3.1. CTC detection
improved biomediator (Fig. 3Bb) lowered the CV from 21.43% to 9.2%, CTCs are important indicators of progression in various cancers. To
indicating that modifying the five chips with the improved biomediator simulate detection of CTCs in patients with cancer, blood samples were
significantly improved consistency. Likewise, in chips with the original obtained from a healthy individual and a patient with lung cancer from
biomediator, using the calibrated settings (Fig. 3Bc) instead of the Shuang Ho Hospital. Flow cytometry performed at the hospital indicated
original settings (Fig. 3Ba) improved the CV from 21.43% to 7.22%, the presence of cytokeratin 7 (CK7), a biomarker associated with lung
again demonstrating improved consistency. Combining the calibrated cancer CTCs, in the blood sample of the lung cancer patient. CK7 was
settings and modified biomediator resulted in the most substantial therefore chosen as the detection target, and a biosensor was con­
improvement to consistency between chips, lowering the CV to 1.65%, structed using an anti-CK7 antibody-modified chip for detection. The
as shown in Fig. 3Bd. In addition, consistency between production detection results (Fig. 6Ab) deduced from the calibration curve
batches was confirmed (Fig. 3C). EIS of three different batches of demonstrated a detection sensitivity of 2 target cells in 1 mL solution
modified chips with improved components yielded a CV of 1.82%, (Fig. 6Aa), indicating functional CTC biomarker detection.
suggesting that calibration of the production settings and the improved
biomaterials considerably elevated consistency. The functionality and 3.3.2. DNA/RNA detection
reproducibility of the final product were then evaluated by using a fully DNA/RNA concentrations are often examined to monitor changes in
constructed biosensor to detect cTnI (Fig. 3D). Fig. 3Da indicates that gene expression in cells. For example, YKL 40 is an inflammatory factor
increased antigen concentrations elicited stronger impedance signals, in many cancers and may indicate changes in gene expression or even
thereby verifying functionality. Biosensors were then constructed using cancer cell transition/migration (Jefri et al., 2015). Therefore, to mimic
chips from different batches to demonstrate post-modification consis­ realistic POC settings, anti-YKL-40 DNA probe-modified chips were used
tency across the production process. High reproducibility was achieved,
to detect YKL-40 fragments in total RNA of different lung cancer cell
regardless of batch origination (Fig. 3Db). Overall, the combination of lines. The detection results (Fig. 6Bb) deduced from the calibration
optimized SMT production settings and the improved biomediator
curve revealed a limit of detection (LOD) of 1 fM (Fig. 6Ba). High
greatly enhanced the consistency of the modified chips, resulting in readings were detected in H1975 cells, and low readings were detected
significantly improved detection reproducibility.
in H292 and HCC827 cells, consistent with the results of PCR.

3.2. Accuracy and stability 3.3.3. Curcumin detection

To determine the accuracy of the improved biosensor, cTnI detection
in whole blood was assessed. Whole blood samples were collected and
analyzed using both the improved biosensor in the laboratory and the
gold standard method in the hospital. Similar results were observed
between the SMEB platform and the hospital method (Fig. 4). In addi­
tion, the SMEB results showed low relative deviation from the hospital
results; specifically, relative deviation was limited to 7% across all
samples (Table 1), indicating a minimum accuracy of 93%. Overall,
these results suggest that biosensors produced with the SMEB platform
show strong potential to maintain accuracy and high clinical feasibility
once integrated into POC applications.
Assessment of the concentrations of functional compound is critical
to some types of agriculture. For example, curcumin is important in
turmeric cultivation. Therefore, detection of curcumin dissolved in
ethanol was performed using αS1-casein-modified chips (αS1-casein has
a high binding affinity for curcumin (Sneharani et al., 2009)). The
detection results (Fig. 6Cb) deduced from the calibration curve sug­
gested the LOD was 10 pM (Fig. 6Ca), indicating functional curcumin
detection. These results were consistent with those obtained using
high-performance liquid chromatography (HPLC).

4
L.-C. Chen et al. Biosensors and Bioelectronics 155 (2020) 112111

3.4. Preparation for POC use 4. Discussion

First, a SOP was defined to optimize SAM formation and bioreceptor
modification for highly accurate and consistent detection signals in each
production batch. The reaction time was then investigated; when the
antigen concentration was greater than 100 pg/mL, only 10 min of re­
action time were required for accurate detection (Fig. 7A). Next, to
simplify measurement, the sensitivity of various measurement fre­
quencies was tested. A frequency of 0.1 Hz yielded the highest sensi­
tivity (Fig. 7B) and was deemed optimal. Finally, detection using a
biosensor produced with the SMEB platform was demonstrated in a
video recording.
Supplementary video related to this article can be found at htt
ps://doi.org/10.1016/j.bios.2020.112111.
Numerous factors can seriously affect the reproducibility, accuracy,
and stability of biosensors, and have resulted in much attention in the
fields of sensing towards optimizing biosensors. Most recent platforms
achieved improved reproducibility, accuracy, or stability individually
by adjusting either the electrodes or surface materials via different
methods of biosensor production. These developments include
enhanced accuracy through integration of highly vertical ZnO nanorods
(Shrivastava et al., 2017), incorporation of covalent organic framework
(COF) materials into sensing fields (Liu et al., 2019), use of fluorescence
(Lee et al., 2017), and application of electrochemical immunodetection
(Yang et al., 2019), as well as superior reproducibility through use of a
platinum network as a counter electrode (Zhao et al., 2018), utilization
of sulfur nitrogen co-doped ordered mesoporous carbon (SN-OMC) and
thymine Hg2þ thymine (T Hg2þ T) mismatch structure as

Fig. 3. Reproducible SMEB platform

establishment. (A) Consistency analysis
using SEM and EIS of bare chips pro­
duced with original and improved set­
tings. (B) SEM and EIS consistency
analysis of modified chips produced
with original settings þ original bio­
mediator (a), original settings þ
improved biomediator (b), improved
settings þ original biomediator (c), and
improved settings þ improved bio­
mediator (d). (C) EIS consistency anal­
ysis of different batches produced with
the improved settings þ improved
biomediator-modified chips. (D) Func­
tional verification of anti-cTnI antibody-
modified chips and cTnI antigen based
on measuring EIS in different settings þ
biomediator combinations (a) and in
different batches (b).

5
L.-C. Chen et al.
Fig. 4. Examination of the accuracy of the SMEB platform using anti-cTnI antibody-modified chips to detect cTnI antigen in whole blood samples. The actual
concentrations were verified using the gold standard method in the hospital. (For interpretation of the references to colour in this figure legend, the reader is referred
to the Web version of this article.)
electrochemical aptasensors (Lai et al., 2018), and development of
imprinted molecular sensors (Fu et al., 2019). However, POC guidelines
demand high performance in all three aspects (reproducibility, accu­
racy, and stability).
This research focused on resolving the limiting drawbacks of two
highly advantageous production techniques—SMT-produced electrodes
and streptavidin biomediators—to provide excellent reproducibility,
accuracy, and stability at the same time. The combination of SMT pro­
duction settings optimized for label-free affinity detection (Fig. 1) and a
biomediator produced with a linker of ideal flexibility and rigidity
(Fig. 2A) was investigated as a method of developing an electrochemical
biosensor that possesses reproducibility, accuracy, and stability suitable
for POC use. Calibrating electrode thickness to greater than 0.1 μm and
surface roughness to less than 0.3 μm allowed our chips to perform
better label-free affinity detection, and also improved the consistency of
electrode assembly (Fig. 3). Previous studies demonstrated that current
production settings do not fall in these ranges and have lower detecting
consistency (Butterworth et al., 2019; Hierlemann et al., 2003). Inclu­
sion of a GW linker with both flexibility and rigidity (Lower et al., 2005)
to improve the biomediator ensures that the bioreceptor is properly
immobilized and that its function is not hindered by limited orientation
(Caroselli et al., 2018; Trilling et al., 2013), while maintaining enough
distance between the receptor domain and chip to preserve function
(Chen et al., 2013; Li et al., 2016).
To validate these improvements, several experiments involving
detection of cTnI were conducted to assess various aspects of perfor­
mance. First, reproducibility was investigated. Fig. 3A and Fig. 3Bc show
the improved consistency from calibrated settings in bare and modified
chips, respectively. Fig. 3Bb shows the increased consistency achieved
through uniform immobilization of the bioreceptor facilitated by the
improved biomediator. Comparison between the SEM images in Fig. 3A
as well as between those in Fig. 3B show no visible differences in
commonly modified parameters, such as gap etching size, suggesting
Fig. 5. Stability analysis of the SMEB platform during storage at RT for nine weeks, using anti-cTnI antibody-modified chips to detect cTnI antigen by EIS every week.
6
Biosensors and Bioelectronics 155 (2020) 112111
that the decreased CV in the modified chips resulted directly from
optimized calibration of the settings and unique biomediator, respec­
tively. Fig. 3Bd shows the consistency of chips modified with both im­
provements, while Fig. 3C and Fig. 3D show high consistency between
batches and high detection reproducibility, respectively. While the
modifications improved chip consistency individually, the combination
of both optimized settings and unique biomediator was essential to
achieving reproducibility that meets POC guidelines.
Accuracy was investigated next (Fig. 4), with the comparison be­
tween detection by the proposed biosensor and hospital tests revealing
low relative deviation and high alignment with the gold standard
method performed by the hospital, and thus high accuracy. This is likely
due to the modified electrode thickness and surface roughness
increasing conductivity (Butterworth et al., 2019; Hierlemann et al.,
2003) and generally complementing label-free affinity detection.
Finally, the stability of the proposed biosensor was evaluated by
assessing signal maintenance over a period of nine weeks. Fig. 5 shows
variation over this period was limited to 6%, suggesting stability suit­
able for POC use.
Optimization of SMT production settings, combined with a unique
biomediator, allowed the development of a biosensor with greater
reproducibility, accuracy, and stability than current products achieve.
Importantly, the CVs of all three of these qualities were improved to less
than 10%, in accordance with the guidelines determined by the CLSI
(EP05-A3, EP24-A2, EP25-A). Using this platform, various biosensors
were constructed to assess their ability to detect common POC targets
including CTCs, DNA/RNA, and curcumin (Fig. 6). In terms of CTC
detection, the biosensors demonstrated a detection limit of 2 cells/mL
(Fig. 6A), which is far superior to the detection limit of 52.24 cells/mL in
a recent study (Chen et al., 2019). For nucleic acid detection, this study
achieved a LOD of 1 fM (Fig. 6B), which is more sensitive than recently
reported detection at picomolar level (Aoki et al., 2019). Lastly, our
biosensors showed a LOD of 10 pM for curcumin (Fig. 6C), while a
L.-C. Chen et al. Biosensors and Bioelectronics 155 (2020) 112111

Fig. 6. SMEB platform applications. (A)

Detection of circulating tumor cells
(CTCs) using anti-CK7 antibody-modi­
fied chips and EIS. Calibration curve (a)
and results for whole blood samples
from a healthy individual and a patient
with lung cancer (b). (B) DNA/RNA
detection using anti-YKL-40 DNA probe
(5’-TTAGGGTGGTAAAATGCTGTT-3’)-
modified chips, YKL-40 DNA fragment
(5’-AACAGCATTTTACCACCCTAA-3’),
and measurement by DPV. Calibration
curve (a) and total RNA (H1975, H292
and HCC827 cell lines) detection (b);
the results were consistent with PCR
analysis (b). (C) Curcumin detection
using αS1-casein-modified chips and
DPV. Calibration curve (a) and curcu­
min detection from normal leaves and
turmeric leaves (b); the results were
consistent with HPLC analysis (b).

Fig. 7. Optimization for POC use. (A) Reaction time investigation using anti-cTnI antibody-modified chips to detect cTnI antigen after different antigen reaction
times (from 10 min to 60 min). (B) Measurement simplification from broad frequency scanning (EIS) to fixed frequency scanning (IMPT). The tested frequencies
included 0.1, 1, 10 and 100 Hz at an amplitude of 10 mV.

previous study indicated a curcumin detection limit of 0.109 μM in
human blood serum (Zokhtareh and Rahimnejad 2018). The proposed
biosensors made with improved electrodes and biomediator were able to
break through the current limits of detection for CTCs, DNA/RNA, and
curcumin and may prove useful in the development of various products
designed to detect common POC targets. Finally, to prepare the platform
for POC use, we investigated reaction time (Fig. 7A), simplified the re­
action measurement (Fig. 7B), miniaturized the device, and recorded a
demonstration video to ensure quick detection, portability, and ease of
use, three key characteristics of POC products (Baryeh et al., 2017).
Through this research, we have developed a platform capable of pro­
ducing biosensors that possess exceptional reproducibility, accuracy,
and stability, and show exceptional promise for successful integration in
POC applications.
and surface roughness to less than 0.3 μm. In addition, the streptavidin
biomediator was improved through addition of a GW linker, which
provides both ideal flexibility and rigidity for bioreceptor immobiliza­
tion and thus performs better than previously used linkers. These im­
provements considerably enhanced performance, allowing all three
required characteristics (reproducibility, accuracy, and stability) to
meet POC standards. To demonstrate the wide range of the platform,
biosensors were produced to prove functionality for detection of CTCs,
DNA/RNA, and small compounds, three common types of POC targets.
Finally, steps were taken to optimize the platform for future use in
realistic applications, by defining a SOP for mass production and
ensuring storage stability. In summary, this SMEB platform simulta­
neously achieves reproducibility, accuracy, and stability and is suitable
for integration into a wide variety of POC products.

5. Conclusion Declaration of competing interest

This project aimed to establish a platform capable of producing The authors declare that they have no known competing financial
biosensors with reproducibility, accuracy, and stability that meet the interests or personal relationships that could have appeared to influence
POC guidelines set by the CLSI. The settings for SMT production of the the work reported in this paper.
electrodes were optimized, calibrating thickness to greater than 0.1 μm

7
L.-C. Chen et al.
CRediT authorship contribution statement
Lung-Chieh Chen: Conceptualization, Methodology, Investigation,
Writing - original draft, Visualization. Erick Wang: Writing - original
draft, Writing - review & editing. Chun-San Tai: Investigation, Writing -
review & editing, Visualization. Yuan-Chen Chiu: Investigation, Visu­
alization. Chang-Wei Li: Resources. Yan-Ren Lin: Resources. Tsung-
Han Lee: Resources. Ching-Wen Huang: Resources. Jung-Chih Chen:
Resources, Methodology. Wen Liang Chen: Conceptualization, Meth­
odology, Visualization, Writing - review & editing, Supervision.
Acknowledgements
This work was supported in part by the Green and Intelligent Agri­
cultural System and the Spatiotemporal Control of Epithelial to
Mesenchymal Transition by Tantalum Oxide Artificial Microenviron­
ments: The Next Step towards Personalized Medicine sponsored by
Taiwan Ministry of Science and Technology [grant numbers MOST 106-
2314-B-009-001, 106-3114-8-009-004, and 107-2823-8-009-004]; and
in part by the Novel Bioengineering and Technological Approaches to
Solve Two Major Health Problems in Taiwan sponsored by the Taiwan
Ministry of Science and Technology Academic Excellence Program
[grant number MOST 108-2633-B-009-001].
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.bios.2020.112111.
Funding
<Ministry of Science and Technology>
Grant Number: MOST 106-2314-B-009-001, 106-3114-8-009-004,
107-2823-8-009-004, 108-2633-B-009-001.gs5
References
Anik, Ü., 2017. Electrochemical medical biosensors for POC applications. Medical
Biosensors for Point of Care (POC) Applications 275–292.
Aoki, H., Corn, R.M., Matthews, B., 2019. MicroRNA detection on microsensor arrays by
SPR imaging measurements with enzymatic signal enhancement. Biosens.
Bioelectron. 142, 111565.
Baryeh, K., Takalkar, S., Lund, M., Liu, G., 2017. Introduction to medical biosensors for
point of care applications. Medical Biosensors for Point of Care (POC) Applications
3–25.
Butterworth, A., Blues, E., Williamson, P., Cardona, M., Gray, L., Corrigan, D.K., 2019.
SAM composition and electrode roughness affect performance of a DNA biosensor for
antibiotic resistance. Biosensors (Basel) 9 (1).
Caroselli, R., Garcia Castello, J., Escorihuela, J., Banuls, M.J., Maquieira, A., Garcia-
Ruperez, J., 2018. Experimental study of the oriented immobilization of antibodies
on photonic sensing structures by using protein A as an intermediate layer. Sensors
(Basel) 18 (4).
Chen, X., Zaro, J.L., Shen, W.C., 2013. Fusion protein linkers: property, design and
functionality. Adv. Drug Deliv. Rev. 65 (10), 1357–1369.
Chen, Y., Peng, J., Lai, Y., Wu, B., Sun, L., Weng, J., 2019. Ultrasensitive label-free
detection of circulating tumor cells using conductivity matching of two-dimensional
semiconductor with cancer cell. Biosens. Bioelectron. 142, 111520.
da Silva, E.T.S.G., Souto, D.E.P., Barragan, J.T.C., Giarola, J.de F., deMoraes, A.C.M.,
Kubota, L.T., 2017. Electrochemical biosensors in point-of-care devices: recent
advances and future trends. ChemElectroChem 4 (4), 778–794.
Daniels, J.S., Pourmand, N., 2007. Label-free impedance biosensors: opportunities and
challenges. Electroanalysis 19 (12), 1239–1257.
Biosensors and Bioelectronics 155 (2020) 112111
Dundas, C.M., Demonte, D., Park, S., 2013. Streptavidin-biotin technology:
improvements and innovations in chemical and biological applications. Appl.
Microbiol. Biotechnol. 97 (21), 9343–9353.
Fu, K., Zhang, R., He, J., Bai, H., Zhang, G., 2019. Sensitive detection of ketamine with an
electrochemical sensor based on UV-induced polymerized molecularly imprinted
membranes at graphene and MOFs modified electrode. Biosens. Bioelectron. 143,
111636.
Hierlemann, A., Brand, O., Hagleitner, C., Baltes, H., 2003. Microfabrication techniques
for chemical/biosensors. Proc. IEEE 91 (6), 839–863.
Ikonomova, S.P., Le, M.T., Kalla, N., Karlsson, A.J., 2018. Effect of linkers on
immobilization of scFvs with biotin-streptavidin interaction. Biotechnol. Appl.
Biochem. 65 (4), 580–585.
Jefri, M., Huang, Y.N., Huang, W.C., Tai, C.S., Chen, W.L., 2015. YKL-40 regulated
epithelial-mesenchymal transition and migration/invasion enhancement in non-
small cell lung cancer. BMC Canc. 15, 590.
Johnson, S., Krauss, T.F., 2017. Label-free affinity biosensor arrays: novel technology for
molecular diagnostics. Expet Rev. Med. Dev. 14 (3), 177–179.
Lai, C., Liu, S., Zhang, C., Zeng, G., Huang, D., Qin, L., Liu, X., Yi, H., Wang, R., Huang, F.,
Li, B., Hu, T., 2018. Electrochemical aptasensor based on sulfur-nitrogen codoped
ordered mesoporous carbon and thymine-Hg(2þ)-thymine mismatch structure for
Hg(2þ) detection. ACS Sens. 3 (12), 2566–2573.
Lee, W.I., Shrivastava, S., Duy, L.T., Yeong Kim, B., Son, Y.M., Lee, N.E., 2017.
A smartphone imaging-based label-free and dual-wavelength fluorescent biosensor
with high sensitivity and accuracy. Biosens. Bioelectron. 94, 643–650.
Li, G., Huang, Z., Zhang, C., Dong, B.J., Guo, R.H., Yue, H.W., Yan, L.T., Xing, X.H., 2016.
Construction of a linker library with widely controllable flexibility for fusion protein
design. Appl. Microbiol. Biotechnol. 100 (1), 215–225.
Liu, X., Huang, D., Lai, C., Zeng, G., Qin, L., Wang, H., Yi, H., Li, B., Liu, S., Zhang, M.,
Deng, R., Fu, Y., Li, L., Xue, W., Chen, S., 2019. Recent advances in covalent organic
frameworks (COFs) as a smart sensing material. Chem. Soc. Rev. 48 (20),
5266–5302.
Lower, B.H., Yongsunthon, R., Vellano 3rd, F.P., Lower, S.K., 2005. Simultaneous force
and fluorescence measurements of a protein that forms a bond between a living
bacterium and a solid surface. J. Bacteriol. 187 (6), 2127–2137.
Minamiki, T.M., T, Sasaki, Y., Kurita, R., Niwa, O., Wakida, S., Tokito, S., 2015. An
organic field-effect transistor with an extended-gate electrode capable of detecting
human immunoglobulin A. Anal. Sci. 31, 725–728.
Palazon, F., Benavides, C.M., Leonard, D., Souteyrand, E., Chevolot, Y., Cloarec, J.P.,
2014. Carbodiimide/NHS derivatization of COOH-terminated SAMs: activation or
byproduct formation? Langmuir 30 (16), 4545–4550.
Pohanka, M., 2017. Biosensors based on semiconductors, a review. International Journal
of Electrochemical Science 6611–6621.
Shrivastava, S., Triet, N.M., Son, Y.M., Lee, W.I., Lee, N.E., 2017. Seesawed fluorescence
nano-aptasensor based on highly vertical ZnO nanorods and three-dimensional
quantitative fluorescence imaging for enhanced detection accuracy of ATP. Biosens.
Bioelectron. 90, 450–458.
Siontorou, C.G., Batzias, F.A., 2010. Innovation in biotechnology: moving from academic
research to product development–the case of biosensors. Crit. Rev. Biotechnol. 30
(2), 79–98.
Sneharani, A.H., Singh, S.A., Appu Rao, A.G., 2009. Interaction of alphaS1-casein with
curcumin and its biological implications. J. Agric. Food Chem. 57 (21),
10386–10391.
Stettner, J., Frank, P., Griesser, T., Trimmel, G., Schennach, R., Gilli, E., Winkler, A.,
2009. A study on the formation and thermal stability of 11-MUA SAMs on Au(111)/
mica and on polycrystalline gold foils. Langmuir 25, 1427–1433.
Szunerits, S., Mishyn, V., Grabowska, I., Boukherroub, R., 2019. Electrochemical
cardiovascular platforms: current state of the art and beyond. Biosens. Bioelectron.
131, 287–298.
Trilling, A.K., Beekwilder, J., Zuilhof, H., 2013. Antibody orientation on biosensor
surfaces: a minireview. Analyst 138 (6), 1619–1627.
Tsai, T.C., Liu, C.W., Wu, Y.C., Ondevilla, N.A.P., Osawa, M., Chang, H.C., 2019. In situ
study of EDC/NHS immobilization on gold surface based on attenuated total
reflection surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS).
Colloids Surf. B Biointerfaces 175, 300–305.
Yang, T., Ren, X., Yang, M., Li, X., He, K., Rao, A., Wan, Y., Yang, H., Wang, S., Luo, Z.,
2019. A highly sensitive label-free electrochemical immunosensor based on poly
(indole-5-carboxylicacid) with ultra-high redox stability. Biosens. Bioelectron. 141,
111406.
Zhao, L., Han, H., Ma, Z., 2018. Improved screen-printed carbon electrode for
multiplexed label-free amperometric immuniosensor: addressing its conductivity
and reproducibility challenges. Biosens. Bioelectron. 101, 304–310.
Zokhtareh, R., Rahimnejad, M., 2018. A novel sensitive electrochemical sensor based on
nickel chloride solution modified glassy carbon electrode for curcumin
determination. Electroanalysis 30 (5), 921–927.